Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/pushary/pushary-skill/pushary-chatgptnpx skills add Pushary/pushary-skill --skill pushary-chatgptgit clone --depth 1 https://github.com/Pushary/pushary-skillWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/pushary/pushary-skill/pushary-chatgpt)<a href="https://agentmods.dev/skills/pushary/pushary-skill/pushary-chatgpt"><img src="https://agentmods.dev/badge/skills/pushary/pushary-skill/pushary-chatgpt.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00163 | $0.01982 |
| Opus 5 | $0.00081 | $0.00991 |
| Sonnet 5 | $0.00033 | $0.00396 |
| Haiku 4.5 | $0.00016 | $0.00198 |
Grade A, and why
pushary-chatgpt scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 137 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Pushary: plan first, ask once, always report back
Pushary reaches the person on their phone. They answer from the lock screen or from the dashboard, and either way the question and the answer are recorded.
That changes how you should work. You no longer have to hold a task open in the chat hoping the user is still reading, and you no longer have to guess at a fork because asking would stall you. Work the loop below. It exists to make the number of interruptions small and the number of unrecorded guesses zero.
The loop
- Plan the work and find the decisions before you start.
- Put the plan to the user as one question.
- Ask through
ask_userat every real fork. Never guess. - Finish with
send_notification.
1. Plan before you act
Before the first action, write out the steps in order and mark the points where you would have to choose.
Finding the forks now is the whole efficiency gain. A fork you find while planning can be folded into one question with the others. A fork you find halfway through costs its own interruption, and interruptions are the expensive part.
While planning, sort every open point into one of three piles:
- You can answer it. It is in the request, in the conversation, or derivable from a tool you already have. Answer it and move on. This is not a decision.
- It only matters if a later step goes a certain way. Leave it. Ask when you get there, if you get there.
- The user has to answer it. Carry it to step 2 and ask it with the plan.
2. Put the plan to the user once
One ask_user call with type: "confirm".
question: one line, answerable at a glance. "Start on this plan?"context: the numbered steps and any assumption you made. Under 500 characters, so this is the plan, not an essay.intent: the user's own request, one line.action: what you will actually do first.blocker: why you stopped here. For a plan: "Approving once means I will not stop again unless something is irreversible."
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 5d ago First seen · 137 lines · 163 tokens per session scan A 5bb0f0e26ea3
pushary-chatgpt is a skill published in the GitHub repository Pushary/pushary-skill (1 stars, last pushed 7d ago), licensed MIT. It adds 163 tokens to every session and 1,982 once invoked, about $0.0008 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
save-state
Saves current working context to production/session-state/active.md AND writes an atomic per-task checkpoint to .tasks/checkpoints/[taskid].md. Run before any major context reset, when context usage exceeds 60%, or when ending a work session.
resume-from
Restores cognitive state from an atomic checkpoint in .tasks/checkpoints/[taskid].md, enabling instant recovery at the exact point of failure without re-running the full task.
start
Guides first-time onboarding by asking where the user is in their project journey, then routing them to the right workflow. Use at the beginning of a new session without context, or when the user runs /start for the first time.
inbox-triage
Processes a batch of emails — categorizes them, drafts responses where appropriate, and tells the user which ones need a human decision. Use when the user asks to triage the inbox, clear email, batch-process email, or prep email responses. Best run once or twice a day, not continuously.
meeting-debrief
Turns raw meeting notes or a rough recap into structured action items, a follow-up email draft, and a saved record for future reference. Use when the user says they just finished a meeting, wants to debrief a call, needs to send a follow-up, or wants to capture notes from a meeting. Pairs with meeting-prep — saves to…
meeting-prep
Produces a one-page brief before a meeting — who's attending, what's been discussed before, what to expect, what to ask. Pulls from calendar, email threads with the attendees, any prior meeting notes, and project context. Use when the user asks to prep for a meeting, get ready for a call, or brief me on an upcoming…